Interesting about the bird tracking, thanks for sharing. While that establishes that a bird can be tracked by a hobbyist sport radar system at 6 miles, it doesn't establish anything about the ship's radar interaction with birds because 1) they're different systems and 2) they're designed for different purposes. The interview with a radar operator on one of the ships (think it was not go-fast but same system or similar) said it filters out birds and similar clutter (weather effects too most of the time, he describes what those look like as well), so to me it seems most likely they aren't seeing bird targets. Is it possible they are anyway? Maybe. Is it also possible that a modern radar intended for tracking militarily significant contacts filters non-militarily significant contacts like birds and so forth out? Yes, and we have testimony to that effect. So while I can't claim to know for sure either way, we have speculation on one case with a different device with a different purpose and positive testimony on the relevant system in the other, so I'd have to put more weight on the "ship radar doesn't show birds to the radar operator" possibility whether it can detect and/or track birds or not. Might they get a ping occasionally? Maybe. Are they likely to send jet fighters to intercept a bird? Well..... Like I said before, if our carrier groups are sending jets to intercept birds and the pilots are baffled by what they're seeing upon arrival (especially when there's visual contact off FLIR), we have much more serious problems than UFOs.
Jet IR/radar slave quote from the manual: That says the IR can be slaved to a radar and point the IR system toward a locked radar target, but it doesn't say the reverse is true: That the radar can be swiveled to lock an IR target. Perhaps it says it can somewhere else in the manual? Even if it's not in the DCS manual, it's possible that the real F18 can do it even if the DCS F18 can't (DCS isn't a perfect recreation after all, and if it's classified it better not be in DCS to begin with). If it could I'd think the manual would mention it.
What seems pretty clear is the radar didn't see the thing when it was just a couple of miles away. If it did, they could have locked the radar to it and slaved the FLIR and saved themselves the trouble of hunting around for it manually. If the ship didn't have a radar contact on this one like the other two did, how did the pilots run across this in the first place? By chance? If it's a bird and the F18 radar can see birds, why can't the radar see it? Sounds like we're trying to have it both ways at the same time. Chances are they ran across this not by chance, but the same way the pilots for the other two videos did: Ship radar tracked them and vectored them to intercept. Do I know that for sure in the go fast case? Nope.
Ironically enough the "no target on radar" would be a point for the bird hypothesis, but then we couldn't argue that the radar locked it and provided range information. Either the plane's radar can see the target or it can't. On at least one of the other videos the ship radar vectored the jets out, jets got a ping on their own radars, and in Underwood's case he locked the target briefly and the thing jammed his radar. If that was really a plane as many argue, why did the radar fail to identify the target? That would be an unidentified aircraft the computer doesn't know about. If an airplane was flying around up there and jammed Underwood's radar, that's an act of war according to Fravor, so there'd probably be a phone call to the president over it and some serious trouble. So in tic/tac or gimble (I forget which one was Underwood) we have an aircraft that neither the ship's radar or either of the two F18s radars could identify, plus we have a jamming event. To my mind that doesn't fit with "small civilian aircraft running no transponder that lost its way."
Either way we still don't know for sure if the range info is accurate on go fast. I'm leaning toward "no," which is what two out two pilots have said about this. The symbology for laser tracking is missing so we can be pretty sure there's no laser/IR beam on the target, and both pilots say the laser wouldn't be used in this situation anyway. If it's true that the range info is wrong, then there's likely no radar lock either and perhaps the IR/Radar slave mode only works in one direction (radar can put IR sensor on target, but IR sensor can not put radar on target. Manual mentions one but not the other AFAIK unless you've seen otherwise).
On the FLIR look-at angle ranging: I agree with you, geometrically it doesn't make sense that they could determine range from that. I certainly can't from angles alone unless you have two separate aircraft reasonably far apart tracking the object simultaneously so you can triangulate the position. Pilots usually get explanations just deep enough to operate the systems effectively, but that's about it. They don't know the systems anywhere near as well as the engineers do. If I was hired to write targeting pod software to do this, the first thing I'd try is image processing to determine angular size while trying to match the image against a 3D model at all different rotations with AI. If identification probability was too low, I'd call it "unidentified" and try to provide range information by assuming a standard fighter plane size and display that to the pilot. I imagine having some sort of rough range estimate is preferable to nothing.
I think in the FLIR go fast case, that's probably what we're seeing: An estimate derived something along those lines. It's very hard to be accurate with that though, and as you know already, if the actual size isn't the same as the default size, the range info could be off by a long way. If it's half the size the software is guessing it is, the range would be off by 2 miles one way, and if it's double, it'd be off by 4 miles the other, which throws all the analyses (including mine) out the window for the most part. While they're interesting to do and are often informative when they put upper and lower bounds on unknown data (wind speed, jet turn rate, target speed, etc.), we have to take them for what they are:
Each one is one possible solution in a set of infinite possible solutions, nothing more.
This is worth expounding on in detail and ties into confirmation bias on all sides of this issue:
What's bothering me most about all this is people on all sides are simply latching onto whichever solution fits their own wishful thinking and declaring something along the lines of "since analysis X's solution fits explanation X, explanation X is probably right
because it is one possible solution to the problem." While it's true that it's a possible solution to the problem and their conclusion could very well be correct, the conclusion is arrived at via fallacious logic because it ignores the rest of the solution space.
For example, if I do an analysis that puts a ground speed of 50 knots on go-fast, then add 50 knots head wind to target to bring it to rest relative to the wind, I now have one solution in which TAS is 0. Therefore "balloon" is a possible solution. While it's true that "balloon" is in the solution space and might be the correct answer, I can not simply stop at this point and say "therefore it
is a balloon." I can not even say "therefore it is
probably a balloon."
"Probably" equates to "probability." To say "X is probably true" is to claim that there is a higher probability of explanation X being correct than any other explanation. The reason the explanation is not more probable in this example is this: If an unknown 50 knot wind can be inserted in one direction, it can also be inserted in the opposite direction with equal probability because we lack wind information. Now we have a new solution of 100 knots TAS instead of 0. The 100 knot solution is equally probable. We are left with no way to differentiate between "balloon" and "something flying 100 knots TAS into the wind, definitely not a balloon." We need wind data to narrow the solution space enough to draw a definitive conclusion one way or another.
If we are free to simply insert any wind speed we wish, we can go the other direction like I did in my analysis. Jet turn rate could be 0. While I doubt it really is 0, it is possible so therefore the analysis lies within the solution space. So in that analysis I can get 124 knots, and within that I can insert wind speed possibilities as well to generate even more solutions. I don't think anyone would disagree that a 124 knot wind speed difference between those altitudes is out of the realm of possibility, so I can set the wind direction to bring the target air speed to 0 and get a ballon. However, the probability of that wind blowing in the opposite direction is precisely the same, so I could with equal probability say the TAS of the target is 248 knots.
Because the probability of both answers is identical given the available measurements we have access to, we can not accurately say any of the following statements:
1) "TAS is 0 or near 0. It's probably a balloon or bird."
2) "TAS is 248 knots. It's definitely not a balloon or bird"
While they both remain possibilities and one of them may very well be correct given our limited information (not a problem for the Navy, more on that later), we can not say that one is "probably more likely" than the others if the free parameter we're manipulating in the model has an equal probability of facing in either direction and of having any value between 0 and 124 knots or whatever the possible range might be. It could be a headwind or a tail wind. Both are equally probable and therefore we can not pick one solution over the other. The problem gets even worse as you add more free parameters, especially if any have high sensitivity (e.g. turn rate).
Nevertheless, we tend to grab the solution we want to be true and say "since this solution is possible, I know what it "probably" is, so that's good enough for me" and we're done. Nobody on any side of this should be doing that. That's how confirmation bias creeps into analyses like these (including mine) when there is too little information available to draw a definitive conclusion one way or another. I.e., arriving at 100% probability of one solution being correct to another. Then we KNOW the answer. Until then, we don't, so it's unidentified/unknown.
We can go even further than this and use the slower speed analyses that land on 30 knots TAS or so for the target. Stick a 120 knot wind speed difference between the two altitudes into that one and you can make the object fly backwards. So that is also a possible solution and the probability of it being correct is the same as it is for every other solution.
Notice the range of possibilities we can get just using wind speed as the only free parameter. With one parameter I can make it hover, go 250 knots, or even fly backwards. Now add a second or third or fourth free parameter to the model. Add unknown range, unknown turn rate, and whatever else is missing to really know for sure what it is, and I can turn go-fast into almost anything I want it to be.
John von Neumann famously said:
With four parameters I can fit an elephant, and with five I can make him wiggle his trunk.
It gets us no closer to knowing what is "probably right" because we are missing all of the free parameter information necessary to reduce the solution space enough to positively identify the object. Maybe we can establish that range is accurate (not looking good so far) so we can eliminate one free parameter. Maybe we can find winds aloft data in the area and eliminate another. Maybe we can get turn rate telemetry and eliminate another.
Now here's the kicker and why we should take the "unknown" status of these videos rather seriously: Since this video is now officially in their "unknown" bucket, they have done their own analysis. ATIP head who is charged with this said they in fact did do so. They do not have the same free parameter problems the rest of us do. Navy/Pentagon/US intelligence/whatever has (or had, I don't know how long information like that is kept) all of that information because everything that can be recorded on a flight is recorded. They will have the precise temperature of the left rear compressor at 1:02:25.21 into the flight along with every other telemetry output.
Here's what this means for free parameters and why their analysis is so much more significant than mine or anyone else's:
They will have a precise GPS track of the jet and precise turn rate information, so do not need to guess at the flight path like we do. That eliminates turn rate as a free parameter. That narrows the solution space.
They will know the exact status and accuracy of the FLIR range readout and will have whatever radar data is available from not one, but two aircraft and the ship's radar. That eliminates range as a free parameter and narrows the solution space even further.
Remember the ships vectored out the jets with their radars in at least two of these videos, so there is ship radar information which we don't see. So they'll have precise radar tracks on at least one or two of the objects which can be cross correlated with the information on the jet (FLIR pod and so on), multiple sources for the same variable, so even the ground speed of the targets is probably known to them at every point long before and after the videos. Corbell's ship radar video demonstrates this to be the case and is confirmed by testimony from one of the ships radar operators involved in one of the three videos. They can simply measure target velocity with ship radar directly. That eliminates target speed information as a free parameter and narrows the solution space even further.
What this means is they have a heck of a lot more information than we do and a lot less free parameters to fiddle with in their analyses than the rest of us have. They can simply measure all the values we are stuck guessing. Their solution space is much narrower as a result. The probabilities of each possible solution being the correct one are narrowed down more than we could ever hope for, so they are able to rule out possibilities when the rest of us can not. If they could determine from their telemetry that the TAS is not 0, then they could have definitively ruled out the balloon hypothesis even before any of us saw the video. I would think they would have categorized the video accordingly. They didn't. They categorized it as "unknown."
Now here's what keeps me up at night: Their conclusion is not "bird" or "balloon" or "airplane" despite having none of the free parameters that we have. To them it's "unknown." I don't know if that means they were unable to decide on what it was or if they were actually able to rule those out of the possibilities. That would be very interesting to know. While it's fun to make these models and give it our best shot, none of us are going to do any better than they did. We really just have to say "we don't know the correct solution for what we see in these videos" and leave it at that. To us out here in internet land, a balloon is one of many possibilities, but to them it apparently wasn't, or they weren't able to determine definitively that it was even with all the telemetry info the rest of us are lacking. We should not be too quick to dismiss that fact and be too confident about our own analyses, thinking we've outsmarted US Intelligence by finding solutions via manipulation of free parameters that they simply don't have to deal with at all in theirs.
So what are these things? The only rational conclusion we can draw given the available information in the videos is "unknown."
).
